Speech-to-text intervention to support text production among students with writing difficulties: a single-case study in nordic countries

Figures & data

Table 1. Participants’ characteristics.

Participants	Country	Age	Gradec	Gender	Nonword reading percentile	Sight word reading percentile	Spelling pre-test percentile	Reading compre- hension pre-test^b	Listening compre- hension pre-test^b
N1	N	9	4	B	4	1	9	0	7
N2	N	12	7	B	17	5	35	7	8
N3	N	12	7	B	2	2	10	6	7
N4	N	10	5	B	1	1	4	2	8
S6^a	S	11	5	B	5	5	5	8	8
S7	S	12	6	G	5	5	15	8	8
S8	S	12	6	G	5	5	5	8	7
S9	S	11	5	B	10	10	10	8	8

N: Norwegian student; S: Swedish student; B: boy; G: girl.

^aThe student’s special needs teacher reported a diagnosed language disorder during the intervention.

^bNumber of correct answers out of 8.

^cIn Sweden, education includes a mandatory preschool class, followed by grade 1 up to grade 9, while in Norway, grades are counted from 1 to 10.

Table 2. Overview of the measurements.

Productivity	Accuracy	Text quality
Total words written	Proportion of words spelled correctly Ratio of correct writing sequences Number of interpunctions and average sentence length	Proportion of vocabulary diversity Proportion of word length (≥seven letters)

Table 3. Procedural fidelity of the intervention.

Participants	Adherence to the manual	STT training	STT technical functioning	Student concentration
N1	52%	100%	92%	52%
N2^a	83%	100%	92%	96%
N3	80%	96%	84%	100%
N4^b	65%	83%	74%	91%
S6	76%	100%	88%	100%
S7	48%	92%	100%	96%
S8	92%	96%	100%	100%
S9	72%	100%	100%	96%

Adherence of the manual represents the percentage of ‘yes’ responses of all 25 sessions.

^a 24 sessions reported for N2.

^b 23 sessions reported for N4.

Figure 1. Total words written. TWW: Total Words written. N4: test 9 is missing. None of the students chose TTS during the intervention’s first test since it was not introduced until session 4. Conditions of used assistive technology: A = STT, B = STT + TTS, C = STT + TTS + Spell-checker, D = STT + TTS + Spell-checker + Word prediction, E = STT + TTS + Word prediction, F = STT + Spell-checker + Word prediction. G = STT + Word prediction, H = STT + Spell-checker.

Line chart illustrating changes in the number of words by individual students across three phases: baseline, intervention, and maintenance. Each student is represented by one of the eight grouped graphs. Students transition from using keyboards in the baseline phase to using speech-to-text in the intervention and maintenance phases, illustrating the impact of the intervention.

Figure 2. Percentage of misspelled words. WSW: Words spelled wrong. Words spelled correct, WSC index = WSC/TWW*100, Misspelled words: SUM(100 – WSC INDEX). N4: test 9 is missing.

Line chart illustrating changes in the percentage of misspelled words by individual students across three phases: baseline, intervention, and maintenance. Each student is represented by one of the eight grouped graphs. Students transition from using keyboards in the baseline phase to using speech-to-text in the intervention and maintenance phases, illustrating the impact of the intervention.

Figure 3. The percentage of correct written sequences of total words. CWS: Correct writing sequences. CWS index = CWS/TWW*100. N4: test 9 is missing.

Line chart illustrating changes in the ratio of correct writing sequences by individual students across three phases: baseline, intervention, and maintenance. Each student is represented by one of the eight grouped graphs. Students transition from using keyboards in the baseline phase to using speech-to-text in the intervention and maintenance phases, illustrating the impact of the intervention.

Figure 4. Number of sentences and average sentence length. The numbers in the graphs represent the sum of sentences according to the students’ punctuation marks. 1 = One paragraph text, with one or no punctuation.

Line chart illustrating changes in the number of sentences and average sentence length by individual students across three phases: baseline, intervention, and maintenance. Each student is represented by one of the eight grouped graphs. Students transition from using keyboards in the baseline phase to using speech-to-text in the intervention and maintenance phases, illustrating the impact of the intervention.

Figure 5. Vocabulary (VD index and WL index). VD: Vocabulary diversity; VD index: VD/•TWW; WL: Word length ≥ 7 letters; WL index: WL/•TWW. N4: test 9 is missing.

Line chart illustrating changes in the vocabulary diversity index and word length index by individual students across three phases: baseline, intervention, and maintenance. Each student is represented by one of the eight grouped graphs. Students shift from using keyboards in the baseline phase to employing speech-to-text in the intervention and maintenance phases, illustrating the impact of the intervention.

Table 4. AT Conditions during the intervention.

Condition					Counts	% of the total
	STT	TTS	Spell-checker	Word prediction
A	x				9	10%
B	x	x			14	15%
C	x	x	x		40	43%
D	x	x	x	x	9	10%
E	x	x		x	2	2%
F	x		x	x	3	3%
G	x			x	1	1%
H	x		x		12	13%

Total intervention sessions = 92; Not reported/missing = 2 sessions. STT = speech-to-text, TTS = text-to-speech.

Table 5. Baseline results.

	Student	TWW	Average sentence length	WSC index	CWS index	VD index	WL index
	N1^a	32	32.0	81	50	3.0	0.00
Max		87	11.0	100	100	5.7	1.11
Mdn	N2	76	9.3	98	97	5.6	0.54
Min		66	7.5	91	76	5.5	0.12
Max		126	23.0	100	98	7.1	0.56
Mdn	N3	115	23.0	99	96	6.4	0.53
Min		115	18.0	98	92	6.3	0.37
Max		45	45.0	74	30	5.2	0.21
Mdn	N4	37	37.0	57	16	4.4	0.16
Min		22	22.0	31	7	3.6	0.00
Max		106	68.0	98	90	6.4	0.86
Mdn	S6	61	36.5	93	76	5.0	0.45
Min		42	24.5	88	67	4.9	0.12
Max		84	20.3	100	108	6.2	1.33
Mdn	S7	81	12.3	100	100	5.7	0.76
Min		74	9.3	99	93	5.6	0.58
Max		173	53.0	99	105	7.1	1.29
Mdn	S8	119	28.8	98	88	6.4	0.78
Min		85	12.4	94	73	5.5	0.52
Max		91	11.7	98	103	6.5	0.66
Mdn	S9	83	9.1	96	101	6.1	0.55
Min		82	8.3	91	88	6.0	0.42

N: Norwegian students; S: Swedish students; Mdn: Median; TWW: Total words written; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW); WL: Word length (index: Number of unique words ≥ 7 letters/•TWW).

^a N1: Only one measurement in baseline.

Table 6. Intervention results.

	Students	TWW	Average sentence length	WSC index	CWS index	VD index	WL index
Max		130	54.0	100	98	6.8	1.19
Mdn	N1	72	12.6	100	90	5.7	0.59
Min		38	7.3	98	74	4.6	0.00
Max		133	16.6	100	108	6.6	1.39
Mdn	N2	91	11.2	100	98	6.3	0.71
Min		71	9.1	98	93	5.2	0.47
Max		131	14.6	100	103	7.5	1.16
Mdn	N3	88	11.0	100	96	6.4	0.77
Min		69	8.2	99	86	6.1	0.32
Max		331	165.5	100	92	8.5	1.18
Mdn	N4	121	11.0	100	80	6.8	0.77
Min		47	9.1	98	73	5.3	0.25
Max		139	64.0	100	96	7.5	1.06
Mdn	S6	83	27.5	100	94	5.6	0.54
Min		52	19.7	99	73	4.6	0.22
Max		562	40.5	100	95	8.6	1.50
Mdn	S7	299	25.9	100	87	7.7	1.23
Min		130	11.8	99	78	6.2	0.70
Max		374	17.0	100	109	8.1	1.52
Mdn	S8	149	13.2	100	101	6.8	1.03
Min		60	8.6	99	89	5.3	0.39
Max		245	25.0	100	106	8.4	1.48
Mdn	S9	148	13.1	100	98	7.5	1.17
Min		80	10.4	99	87	6.5	0.43

N: Norwegian students; S: Swedish students; Mdn: Median; TWW: Total words written; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW); WL: Word length (index: Number of unique words ≥ 7 letters/• TWW).

Table

Overview of the progressive intervention manual.
Learning goal	Instructions	Special attention
The student -knows how to start and set the STT function and dictate words. -has the ability to dictate one sentence at a time with STT. -can independently use STT to produce a story and TTS to revise it.	Demonstrate how to use STT. Demonstrate how to use TTS.^a Provide students with a procedural checklist to prepare and generate text with STT and revise the text by using TTS and STT.^b Give adaptive feedback to support strategy consolidation.	Notice students’ dictation regarding: Pronunciation Speed Technical issues Timing (turning the microphone on and off when dictating). Pay attention to the student’s autonomy regarding the three-step process.^b Pay attention to students’ motivation by providing feedback on their progress and discussing opportunities and possible dilemmas.

STT: Speech-to-text; TTS: Text-to-speech.

^a TTS was introduced in the fourth intervention session.

^b The procedural checklist is presented below.

Table

The checklist of the three-step process permeating the intervention
1) Prepare the technology	2) Speech-to-text/dictate	3) Revise
a. Prepare  the computer/tablet  and microphone  for dictation.	a. Speak the sentence aloud or think it.	a. Listen to one sentence at a time with TTS.
b. Open a document  and save it.	b. Say the sentence into the microphone using STT.	b. Consider if the phrase is meaningful. c. Check capital letters and periods.
c. Check the  microphone and  sound.	c. Say period or type full stop.

STT: Speech-to-text; TTS: Text-to-speech.

Table

Descriptives N1
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	1	13	2	1	13	2	1	13	2	1	13	2	1	13	2	1	13	2
Mean	32	77	59	32.0	19.1	35.6	81	100	99	50	89	83	3.0	5.6	5.6	0.00	0.59	0.41
SD	NA	26	16	NA	15.7	17.5	NA	1	1	NA	8	0	NA	0.6	0.2	NA	0.30	0.24

N: Number of texts; SD: Standard deviation; B: Baseline; I: Intervention; M: Maintenance; TWW: Total words written; SEN: Average sentence length; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW). WL = Word length (index: Number of unique words ≥7 letters /•TWW). NA: Not applicable.

Table

Descriptives N2
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	6	10	2	6	10	2	6	10	2	6	10	2	6	10	2	6	10	2
Mean	77	99	92	9.3	11.7	9.0	97	99	100	94	99	99	5.6	6.3	6.0	0.54	0.84	0.80
SD	8	22	41	1.5	2.4	1.6	3	1	0	9	4	5	0.1	0.5	0.4	0.36	0.31	0.11

N: Number of texts; SD: Standard deviation; B: Baseline; I: Intervention; M: Maintenance; TWW: Total words written; SEN: Average sentence length; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW). WL: Word length (index: Number of unique words ≥ 7 letters/•TWW).

Table

Descriptives N3
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	3	13	2	3	13	2	3	13	2	3	13	2	3	13	2	3	13	2
Mean	119	96	109	21.3	11.2	14.4	99	100	100	95	97	95	6.6	6.6	6.7	0.49	0.77	0.81
SD	6	19	4	2.9	2.2	2.1	1	1	0	3	5	8	0.4	0.4	0.4	0.10	0.23	0.18

N: Number of texts. SD = Standard deviation. B = Baseline. I = Intervention. M = Maintenance. TWW = Total words written, SEN = Average sentence length, WSC = Words spelled correctly (index: WSC/TWW*100), CWS = Correct writing sequences (index: CWS/TWW*100), VD = Vocabulary diversity (index: VD/•TWW). WL = Word length (index: Number of unique words ≥ 7 letters/•TWW).

Table

Descriptives N4
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	6	9	2	6	9	2	6	9	2	6	9	2	6	9	2	6	9	2
Mean	35	144	81	34.7	28.8	19.9	54	99	100	18	81	81	4.5	6.8	6.1	0.12	0.81	0.60
SD	10	101	19	10.0	51.3	16.2	16	1	0	9	7	6	0.6	1.2	0.9	0.10	0.28	0.16

N: Number of texts; SD: Standard deviation; B: Baseline; I: Intervention; M: Maintenance; TWW: Total words written; SEN: Average sentence length; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW). WL: Word length (index: Number of unique words ≥ 7 letters/•TWW).

Table

Descriptives S6
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	6	10	2	6	10	2	6	10	2	6	10	2	6	10	2	6	10	2
Mean	65	90	75	41.4	35.2	74.5	93	100	99	77	90	78	5.3	5.7	5.5	0.46	0.63	0.52
SD	22	30	2	16.5	15.6	2.1	3	0	2	8	8	6	0.6	1.0	0.1	0.26	0.31	0.24

N = Number of texts. SD = Standard deviation. B = Baseline. I = Intervention. M = Maintenance. TWW = Total words written, SEN = Average sentence length, WSC = Words spelled correctly (index: WSC/TWW*100), CWS = Correct writing sequences (index: CWS/TWW*100), VD = Vocabulary diversity (index: VD/•TWW. WL = Word length (index: Number of unique words ≥7 letters /•TWW).

Table

Descriptives S7
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	3	13	2	3	13	2	3	13	2	3	13	2	3	13	2	3	13	2
Mean	80	314	439	13.9	25.8	162.3	100	100	100	100	86	76	5.8	7.7	6.9	0.89	1.17	1.19
SD	5	123	136	5.7	9.5	148.9	1	0	0	8	5	8	0.4	0.7	0.9	0.39	0.27	0.08

N: Number of texts; SD: Standard deviation; B: Baseline; I: Intervention; M: Maintenance; TWW: Total words written; SEN: Average sentence length; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW). WL = Word length (index: Number of unique words ≥7 letters /•TWW).

Table

Descriptives S8
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	6	10	2	6	10	2	6	10	2	6	10	2	6	10	2	6	10	2
Mean	123	180	135	29.8	13.2	13.7	97	100	100	88	98	103	6.3	6.9	6.5	0.82	1.04	0.69
SD	37	94	75	13.8	2.8	4.8	2	0	0	11	8	1	0.6	0.9	0.6	0.26	0.32	0.36

N: Number of texts; SD: Standard deviation; B: Baseline; I: Intervention; M: Maintenance; TWW: Total words written; SEN: Average sentence length; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW). WL = Word length (index: Number of unique words ≥7 letters /•TWW).

Table

Descriptives S9
	TWW B	TWW I	TWW M	SEN B	SEN I	SEN M	WSC B	WSC I	WSC M	CWS B	CWS I	CWS M	VD B	VD I	VD M	WL B	WL I	WL M
N	3	13	2	3	13	2	3	13	2	3	13	2	3	13	2	3	13	2
Mean	85	157	235	9.7	13.8	16.2	95	100	100	97	99	99	6.2	7.4	7.9	0.54	1.03	1.23
SD	5	43	101	1.8	3.6	0.1	4	0	0	8	5	1	0.3	0.6	0.4	0.12	0.33	0.04

N: Number of texts; SD: Standard deviation; B: Baseline; I: Intervention; M: Maintenance; TWW: Total words written; SEN: Average sentence length; WSC: Words spelled correctly (index: WSC/TWW*100); CWS: Correct writing sequences (index: CWS/TWW*100); VD: Vocabulary diversity (index: VD/•TWW). WL = Word length (index: Number of unique words ≥7 letters /•TWW).